DESCRIPTION
This lecture will introduce the continual learning (also called lifelong learning) as one of the emerging areas of deep learning. We will discuss the motivation, objectives and aims for continual learning methodology and applications, as well as the challenges in developing continual learning systems. The latest directions of research will be presented and discussed, including generative models, architecture and memory expanding models. The latest developments in continual learning will be presented. Architectures enabling continual learning models, and their training will be shortly presented. Experimental results when the latest continual learning systems are applied for successively learning image data will be presented and discussed. Limitations of the existing continual learning systems will also be discussed. Directions of future research will be discussed.
DETAILS
Course type: Invited lecture (in person delivery)
Duration: 1 hour
Level: Postgraduate/PhD
Institution of lecturer: University of York
Notes: A set of questions will be provided after the lecture to the participants to the lecture. Answering to the questions will be essay style, no longer than one paragraph/question and will be expected to be provided in 24.
LECTURER
Prof. Adrian G. Bors
Adrian G. Bors is an Associate Professor in the Department of Computer Science at the University of York, UK. He received the Ph.D. degree in informatics from the University of Thessaloniki, Thessaloniki, Greece, in 1999 and the MSc. degree in electronics engineering from the Polytechnic University of Bucharest, Bucharest, Romania, in 1992. In 1999 he joined the Department of Computer Science, Univ. of York, U.K. Dr. Bors was before a Research Scientist at the University of Tampere, Finland, and had visiting academic appointments at the Univ. of California at San Diego (UCSD), the Univ. of Montpellier, France and at the MBZ Univ. of Artificial Intelligence (MBZUAI), Abu Dhabi, UAE. Dr. Bors has authored and co-authored more than 170 research papers, including 41 in journals. He was an Associate Editor of the IEEE Transactions on Image Processing from 2010 to 2014 and the IEEE Transactions on Neural Networks from 2001 to 2009. He was a Co-Guest Editor for a special issue on Machine Vision for the International Journal for Computer Vision in 2018 and the Journal of Pattern Recognition in 2015. His research interests include continual learning and generative AI as well as their applications in computer vision and image processing.